Experimental studies have described the efficacy of specific treatments (antidepressant medication or specific forms of psychotherapy) for depression; but little is known about how treatment styles or outcomes vary by characteristics of patients or practice setting. We propose to use data from two large quasi-experimental studies, the Medical Outcomes Study and the Prospective Payment Quality of Care Study, to examine variations in quality and outcomes of care for depression for different groups of depressed patients (age groups, men versus women, poor and nonpoor, ethnic minority versus white) and for different types of treatment locations (rural versus urban, for-profit versus nonprofit organizations, public versus private). Further, we propose to determine how variations in outcomes correspond with variations in process of care for depressed patients. We will study depressed adult outpatients (elderly and nonelderly) who initially visited a general medical clinician or a mental health specialists using the Medical Outcomes Study data; and we will study elderly inpatients with depression using the Prospective Payment Quality of Care Study. The aspects of treatment studied will include use of outpatient health services, hospital readmission rates, and clinically-detailed measures of the quality of inpatient care. The outcomes studied will include functional status and well-being, satisfaction with care, post-admission mortality, clinical status at discharge, and hospital readmission rates. The analyses will consist of multiple regression analyses which examine the unique associations between patient and practice-level factors with indicators of processes or outcomes, while controlling for initial severity of illness. The proposed investigator team, including a psychiatrist, a mathematician, a statistician, a health policy analyst, and a psychiatric epidemiologist, has extensive experience working with these data bases. In addition, we propose to use expert clinicians to assist in the formulation of specific hypotheses about interpreting findings.